feat: Add character-level alignment for ElevenLabs

[shiragannavar]

Add character-level alignment metadata support to ElevenLabs TTS plugin for accurate lip sync animation in avatar applications.

Changes:

• Add CharacterAlignment dataclass to core TTS module
• Add optional alignment field to SynthesizedAudio
• Update ElevenLabs plugin to capture and forward alignment data
• Add enable_alignment_forwarding parameter (default: True)
• Emit 'alignment_received' events with character timing data
• Add example application demonstrating lip sync integration
• Add comprehensive documentation for client-side integration

The feature is fully backward compatible:

• Existing code without alignment handlers continues to work
• Can be disabled by setting enable_alignment_forwarding=False
• No breaking changes to existing TTS functionality

Resolves lip sync timing requirements for avatar applications by providing:

• Per-character start and end times in seconds
• LiveKit data channel integration for client delivery
• Event-based architecture for flexible consumption

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}